| 1 | // © 2017 and later: Unicode, Inc. and others. | 
|---|
| 2 | // License & terms of use: http://www.unicode.org/copyright.html | 
|---|
| 3 |  | 
|---|
| 4 | #include "unicode/utypes.h" | 
|---|
| 5 |  | 
|---|
| 6 | #if !UCONFIG_NO_FORMATTING | 
|---|
| 7 | #ifndef __NUMBER_AFFIXUTILS_H__ | 
|---|
| 8 | #define __NUMBER_AFFIXUTILS_H__ | 
|---|
| 9 |  | 
|---|
| 10 | #include <cstdint> | 
|---|
| 11 | #include "number_types.h" | 
|---|
| 12 | #include "unicode/stringpiece.h" | 
|---|
| 13 | #include "unicode/unistr.h" | 
|---|
| 14 | #include "formatted_string_builder.h" | 
|---|
| 15 | #include "unicode/uniset.h" | 
|---|
| 16 |  | 
|---|
| 17 | U_NAMESPACE_BEGIN namespace number { | 
|---|
| 18 | namespace impl { | 
|---|
| 19 |  | 
|---|
| 20 | enum AffixPatternState { | 
|---|
| 21 | STATE_BASE = 0, | 
|---|
| 22 | STATE_FIRST_QUOTE = 1, | 
|---|
| 23 | STATE_INSIDE_QUOTE = 2, | 
|---|
| 24 | STATE_AFTER_QUOTE = 3, | 
|---|
| 25 | STATE_FIRST_CURR = 4, | 
|---|
| 26 | STATE_SECOND_CURR = 5, | 
|---|
| 27 | STATE_THIRD_CURR = 6, | 
|---|
| 28 | STATE_FOURTH_CURR = 7, | 
|---|
| 29 | STATE_FIFTH_CURR = 8, | 
|---|
| 30 | STATE_OVERFLOW_CURR = 9 | 
|---|
| 31 | }; | 
|---|
| 32 |  | 
|---|
| 33 | // enum AffixPatternType defined in internals.h | 
|---|
| 34 |  | 
|---|
| 35 | struct AffixTag { | 
|---|
| 36 | int32_t offset; | 
|---|
| 37 | UChar32 codePoint; | 
|---|
| 38 | AffixPatternState state; | 
|---|
| 39 | AffixPatternType type; | 
|---|
| 40 |  | 
|---|
| 41 | AffixTag() | 
|---|
| 42 | : offset(0), state(STATE_BASE) {} | 
|---|
| 43 |  | 
|---|
| 44 | AffixTag(int32_t offset) | 
|---|
| 45 | : offset(offset) {} | 
|---|
| 46 |  | 
|---|
| 47 | AffixTag(int32_t offset, UChar32 codePoint, AffixPatternState state, AffixPatternType type) | 
|---|
| 48 | : offset(offset), codePoint(codePoint), state(state), type(type) {} | 
|---|
| 49 | }; | 
|---|
| 50 |  | 
|---|
| 51 | class TokenConsumer { | 
|---|
| 52 | public: | 
|---|
| 53 | virtual ~TokenConsumer(); | 
|---|
| 54 |  | 
|---|
| 55 | virtual void consumeToken(AffixPatternType type, UChar32 cp, UErrorCode& status) = 0; | 
|---|
| 56 | }; | 
|---|
| 57 |  | 
|---|
| 58 | // Exported as U_I18N_API because it is a base class for other exported types | 
|---|
| 59 | class U_I18N_API SymbolProvider { | 
|---|
| 60 | public: | 
|---|
| 61 | virtual ~SymbolProvider(); | 
|---|
| 62 |  | 
|---|
| 63 | // TODO: Could this be more efficient if it returned by reference? | 
|---|
| 64 | virtual UnicodeString getSymbol(AffixPatternType type) const = 0; | 
|---|
| 65 | }; | 
|---|
| 66 |  | 
|---|
| 67 | /** | 
|---|
| 68 | * Performs manipulations on affix patterns: the prefix and suffix strings associated with a decimal | 
|---|
| 69 | * format pattern. For example: | 
|---|
| 70 | * | 
|---|
| 71 | * <table> | 
|---|
| 72 | * <tr><th>Affix Pattern</th><th>Example Unescaped (Formatted) String</th></tr> | 
|---|
| 73 | * <tr><td>abc</td><td>abc</td></tr> | 
|---|
| 74 | * <tr><td>ab-</td><td>ab−</td></tr> | 
|---|
| 75 | * <tr><td>ab'-'</td><td>ab-</td></tr> | 
|---|
| 76 | * <tr><td>ab''</td><td>ab'</td></tr> | 
|---|
| 77 | * </table> | 
|---|
| 78 | * | 
|---|
| 79 | * To manually iterate over tokens in a literal string, use the following pattern, which is designed | 
|---|
| 80 | * to be efficient. | 
|---|
| 81 | * | 
|---|
| 82 | * <pre> | 
|---|
| 83 | * long tag = 0L; | 
|---|
| 84 | * while (AffixPatternUtils.hasNext(tag, patternString)) { | 
|---|
| 85 | *   tag = AffixPatternUtils.nextToken(tag, patternString); | 
|---|
| 86 | *   int typeOrCp = AffixPatternUtils.getTypeOrCp(tag); | 
|---|
| 87 | *   switch (typeOrCp) { | 
|---|
| 88 | *     case AffixPatternUtils.TYPE_MINUS_SIGN: | 
|---|
| 89 | *       // Current token is a minus sign. | 
|---|
| 90 | *       break; | 
|---|
| 91 | *     case AffixPatternUtils.TYPE_PLUS_SIGN: | 
|---|
| 92 | *       // Current token is a plus sign. | 
|---|
| 93 | *       break; | 
|---|
| 94 | *     case AffixPatternUtils.TYPE_PERCENT: | 
|---|
| 95 | *       // Current token is a percent sign. | 
|---|
| 96 | *       break; | 
|---|
| 97 | *     // ... other types ... | 
|---|
| 98 | *     default: | 
|---|
| 99 | *       // Current token is an arbitrary code point. | 
|---|
| 100 | *       // The variable typeOrCp is the code point. | 
|---|
| 101 | *       break; | 
|---|
| 102 | *   } | 
|---|
| 103 | * } | 
|---|
| 104 | * </pre> | 
|---|
| 105 | */ | 
|---|
| 106 | class U_I18N_API AffixUtils { | 
|---|
| 107 |  | 
|---|
| 108 | public: | 
|---|
| 109 |  | 
|---|
| 110 | /** | 
|---|
| 111 | * Estimates the number of code points present in an unescaped version of the affix pattern string | 
|---|
| 112 | * (one that would be returned by {@link #unescape}), assuming that all interpolated symbols | 
|---|
| 113 | * consume one code point and that currencies consume as many code points as their symbol width. | 
|---|
| 114 | * Used for computing padding width. | 
|---|
| 115 | * | 
|---|
| 116 | * @param patternString The original string whose width will be estimated. | 
|---|
| 117 | * @return The length of the unescaped string. | 
|---|
| 118 | */ | 
|---|
| 119 | static int32_t estimateLength(const UnicodeString& patternString, UErrorCode& status); | 
|---|
| 120 |  | 
|---|
| 121 | /** | 
|---|
| 122 | * Takes a string and escapes (quotes) characters that have special meaning in the affix pattern | 
|---|
| 123 | * syntax. This function does not reverse-lookup symbols. | 
|---|
| 124 | * | 
|---|
| 125 | * <p>Example input: "-$x"; example output: "'-'$x" | 
|---|
| 126 | * | 
|---|
| 127 | * @param input The string to be escaped. | 
|---|
| 128 | * @return The resulting UnicodeString. | 
|---|
| 129 | */ | 
|---|
| 130 | static UnicodeString escape(const UnicodeString& input); | 
|---|
| 131 |  | 
|---|
| 132 | static Field getFieldForType(AffixPatternType type); | 
|---|
| 133 |  | 
|---|
| 134 | /** | 
|---|
| 135 | * Executes the unescape state machine. Replaces the unquoted characters "-", "+", "%", "‰", and | 
|---|
| 136 | * "¤" with the corresponding symbols provided by the {@link SymbolProvider}, and inserts the | 
|---|
| 137 | * result into the FormattedStringBuilder at the requested location. | 
|---|
| 138 | * | 
|---|
| 139 | * <p>Example input: "'-'¤x"; example output: "-$x" | 
|---|
| 140 | * | 
|---|
| 141 | * @param affixPattern The original string to be unescaped. | 
|---|
| 142 | * @param output The FormattedStringBuilder to mutate with the result. | 
|---|
| 143 | * @param position The index into the FormattedStringBuilder to insert the string. | 
|---|
| 144 | * @param provider An object to generate locale symbols. | 
|---|
| 145 | */ | 
|---|
| 146 | static int32_t unescape(const UnicodeString& affixPattern, FormattedStringBuilder& output, | 
|---|
| 147 | int32_t position, const SymbolProvider& provider, Field field, | 
|---|
| 148 | UErrorCode& status); | 
|---|
| 149 |  | 
|---|
| 150 | /** | 
|---|
| 151 | * Sames as {@link #unescape}, but only calculates the code point count.  More efficient than {@link #unescape} | 
|---|
| 152 | * if you only need the length but not the string itself. | 
|---|
| 153 | * | 
|---|
| 154 | * @param affixPattern The original string to be unescaped. | 
|---|
| 155 | * @param provider An object to generate locale symbols. | 
|---|
| 156 | * @return The same return value as if you called {@link #unescape}. | 
|---|
| 157 | */ | 
|---|
| 158 | static int32_t unescapedCodePointCount(const UnicodeString& affixPattern, | 
|---|
| 159 | const SymbolProvider& provider, UErrorCode& status); | 
|---|
| 160 |  | 
|---|
| 161 | /** | 
|---|
| 162 | * Checks whether the given affix pattern contains at least one token of the given type, which is | 
|---|
| 163 | * one of the constants "TYPE_" in {@link AffixPatternUtils}. | 
|---|
| 164 | * | 
|---|
| 165 | * @param affixPattern The affix pattern to check. | 
|---|
| 166 | * @param type The token type. | 
|---|
| 167 | * @return true if the affix pattern contains the given token type; false otherwise. | 
|---|
| 168 | */ | 
|---|
| 169 | static bool containsType(const UnicodeString& affixPattern, AffixPatternType type, UErrorCode& status); | 
|---|
| 170 |  | 
|---|
| 171 | /** | 
|---|
| 172 | * Checks whether the specified affix pattern has any unquoted currency symbols ("¤"). | 
|---|
| 173 | * | 
|---|
| 174 | * @param affixPattern The string to check for currency symbols. | 
|---|
| 175 | * @return true if the literal has at least one unquoted currency symbol; false otherwise. | 
|---|
| 176 | */ | 
|---|
| 177 | static bool hasCurrencySymbols(const UnicodeString& affixPattern, UErrorCode& status); | 
|---|
| 178 |  | 
|---|
| 179 | /** | 
|---|
| 180 | * Replaces all occurrences of tokens with the given type with the given replacement char. | 
|---|
| 181 | * | 
|---|
| 182 | * @param affixPattern The source affix pattern (does not get modified). | 
|---|
| 183 | * @param type The token type. | 
|---|
| 184 | * @param replacementChar The char to substitute in place of chars of the given token type. | 
|---|
| 185 | * @return A string containing the new affix pattern. | 
|---|
| 186 | */ | 
|---|
| 187 | static UnicodeString replaceType(const UnicodeString& affixPattern, AffixPatternType type, | 
|---|
| 188 | char16_t replacementChar, UErrorCode& status); | 
|---|
| 189 |  | 
|---|
| 190 | /** | 
|---|
| 191 | * Returns whether the given affix pattern contains only symbols and ignorables as defined by the | 
|---|
| 192 | * given ignorables set. | 
|---|
| 193 | */ | 
|---|
| 194 | static bool containsOnlySymbolsAndIgnorables(const UnicodeString& affixPattern, | 
|---|
| 195 | const UnicodeSet& ignorables, UErrorCode& status); | 
|---|
| 196 |  | 
|---|
| 197 | /** | 
|---|
| 198 | * Iterates over the affix pattern, calling the TokenConsumer for each token. | 
|---|
| 199 | */ | 
|---|
| 200 | static void iterateWithConsumer(const UnicodeString& affixPattern, TokenConsumer& consumer, | 
|---|
| 201 | UErrorCode& status); | 
|---|
| 202 |  | 
|---|
| 203 | /** | 
|---|
| 204 | * Returns the next token from the affix pattern. | 
|---|
| 205 | * | 
|---|
| 206 | * @param tag A bitmask used for keeping track of state from token to token. The initial value | 
|---|
| 207 | *     should be 0L. | 
|---|
| 208 | * @param patternString The affix pattern. | 
|---|
| 209 | * @return The bitmask tag to pass to the next call of this method to retrieve the following token | 
|---|
| 210 | *     (never negative), or -1 if there were no more tokens in the affix pattern. | 
|---|
| 211 | * @see #hasNext | 
|---|
| 212 | */ | 
|---|
| 213 | static AffixTag nextToken(AffixTag tag, const UnicodeString& patternString, UErrorCode& status); | 
|---|
| 214 |  | 
|---|
| 215 | /** | 
|---|
| 216 | * Returns whether the affix pattern string has any more tokens to be retrieved from a call to | 
|---|
| 217 | * {@link #nextToken}. | 
|---|
| 218 | * | 
|---|
| 219 | * @param tag The bitmask tag of the previous token, as returned by {@link #nextToken}. | 
|---|
| 220 | * @param string The affix pattern. | 
|---|
| 221 | * @return true if there are more tokens to consume; false otherwise. | 
|---|
| 222 | */ | 
|---|
| 223 | static bool hasNext(const AffixTag& tag, const UnicodeString& string); | 
|---|
| 224 |  | 
|---|
| 225 | private: | 
|---|
| 226 | /** | 
|---|
| 227 | * Encodes the given values into a tag struct. | 
|---|
| 228 | * The order of the arguments is consistent with Java, but the order of the stored | 
|---|
| 229 | * fields is not necessarily the same. | 
|---|
| 230 | */ | 
|---|
| 231 | static inline AffixTag makeTag(int32_t offset, AffixPatternType type, AffixPatternState state, | 
|---|
| 232 | UChar32 cp) { | 
|---|
| 233 | return {offset, cp, state, type}; | 
|---|
| 234 | } | 
|---|
| 235 | }; | 
|---|
| 236 |  | 
|---|
| 237 | } // namespace impl | 
|---|
| 238 | } // namespace number | 
|---|
| 239 | U_NAMESPACE_END | 
|---|
| 240 |  | 
|---|
| 241 |  | 
|---|
| 242 | #endif //__NUMBER_AFFIXUTILS_H__ | 
|---|
| 243 |  | 
|---|
| 244 | #endif /* #if !UCONFIG_NO_FORMATTING */ | 
|---|
| 245 |  | 
|---|